How to Compensate for 3D Print Shrinkage

A 3D printed part can measure smaller or larger than the dimension in CAD. The difference is often called shrinkage, although the final result can also be affected by extrusion, cooling, orientation and calibration.

Build a repeatable measurement record

Measure at the same points with the same instrument after a consistent cooling or conditioning interval. Record nominal size, each raw reading, axis, orientation, material, profile and ambient or chamber conditions that matter. Average only comparable readings; a warped corner should be diagnosed rather than hidden inside a mean value.

Print at least a confirming coupon after applying a factor. If the first coupon is 99.2 mm, the corrected coupon is 100.1 mm and the next is 99.6 mm, the process spread may matter more than another decimal place in scale. Fix looseness, temperature instability, inconsistent extrusion or measurement technique before tuning the factor further.

Do not apply one factor to unrelated geometry

A long bar, thick block and thin ring can cool differently even in the same material. Validate the correction on geometry representative of the real part, particularly when X and Y differ or the part is constrained during cooling. Preserve the unscaled source and apply compensation in a controlled derivative or profile so the assumption can be revised.

Compensation is not a substitute for tolerance design. A mating pair still needs intentional clearance, and an undersized hole may require local hole compensation rather than enlarging every external dimension. Use a functional test piece before committing a long or expensive job.

A tight hole does not establish global shrinkage

Illustrative coupon: the 100.00 mm outer span has readings 99.98, 100.00 and 100.02 mm (mean 100.00; range 0.04), while a 10.00 mm hole has readings 9.68, 9.72 and 9.70 mm (mean 9.70; range 0.04). The outer factor is 100/100 = 1. Applying 10/9.70 globally would enlarge the already correct outer span to about 103.09 mm. The hole requires local diagnosis and compensation, not an inferred whole-part factor.

Use separate sets in the per-axis worksheet and hole/pin worksheet. Per-axis aggregation is designed dimension divided by the mean of comparable included readings; it is not an average of unrelated ratios. Leave missing axes unmeasured. When spread approaches the intended adjustment, investigate the measuring method and process stability rather than treating a precise factor as reliable.

Worked per-axis correction

In an illustrative calculation, a coupon is designed as 100.00 × 60.00 mm and is assumed to measure 99.20 × 59.10 mm after cooling and conditioning. The observed factors are 100 / 99.20 = 1.0081 for X and 60 / 59.10 = 1.0152 for Y. A trial part could therefore use about 100.81% X and 101.52% Y scaling, while Z remains unchanged unless separately measured. Print a second coupon to validate; do not infer Z or another material from those two values.

Keep nozzle, profile, orientation, cooling, chamber conditions, material batch, measurement temperature and measuring points fixed. Repeat the print enough to see whether the correction is larger than normal process variation. Record raw dimensions rather than only the chosen percentage so a future profile change can be diagnosed.

Global scaling is inappropriate when outside dimensions are correct but a hole is tight, when first-layer expansion affects only the base, or when a warped part has no stable scale factor. Use feature compensation, first-layer controls or process correction for those cases. Stop and diagnose mechanics if results drift between nominally identical coupons.

Measure a controlled per-axis test

Print a known X, Y and Z test under the same material and profile, then compare designed and measured dimensions. If a 20.00 mm feature measures 19.80 mm, a provisional factor is 20/19.8, about 1.0101. Repeat before applying it broadly: temperature, orientation and part shape can change the result.

The reliable way to compensate is not to copy one universal PLA, PETG or ABS percentage. Print a small test with the same printer, material and profile as the intended part, measure it, and calculate the adjustment from that result.

The 3D Print Shrinkage Calculator keeps the calculation separate for X, Y and Z and gives a compensated model dimension for a target final size.

Print and measure a simple calibration test

Use a part with clearly measurable straight features in all three axes. A 20 mm cube is a simple starting point, though a test that resembles the final part’s size and orientation is more representative. Measure with suitable calipers after the part has cooled and any intended post-processing is done.

1. record the designed test dimension for X, Y and Z;
2. record the printed measurement for each axis;
3. keep the printer, filament, orientation and slicer profile with the measurement;
4. repeat the test if the part is important or the result is surprising.

Calculate the scale factor

The compensation is based on the ratio between the original design and the printed measurement:

Scale factor = designed test dimension ÷ measured print dimension

If a 20.00 mm feature measures 19.80 mm, the scale factor is 20 ÷ 19.8 = 1.0101, or 101.01%. To aim for a 50 mm final feature under the same conditions, model or scale it to about 50.505 mm.

A positive shrinkage percentage means the printed test was smaller than its CAD dimension. A negative percentage means it printed larger. Neither result is automatically a defect: first confirm the measurement and whether it is representative of the final part.

Keep X, Y and Z separate when needed

A single uniform scale is convenient, but it can hide an axis-specific result. Bed adhesion, cooling, layer stacking and part orientation can affect the result differently across the axes. If X and Y agree but Z does not, use the axis-specific values as a starting point and verify with another test before applying them to a critical model.

Shrinkage is not clearance

Whole-model compensation does not make every hole, pin or snap fit correct. Those features also depend on extrusion width, bridging, orientation and the clearance you actually need between mating parts. Use the Tolerance and Clearance Calculator for those dimensions, then validate the final exported model with the STL Dimensions Checker.

Use measured values as a starting point, not a guarantee

A calibration factor belongs to a particular printer, filament, profile and geometry. Changing nozzle settings, material batch, temperature, cooling or orientation can change the outcome. Test important fits and large dimensions again, especially after changing a slicer profile.